214
ABO and P1 Blood Group Antigen Expression and stx Genotype and Outcome
of Childhood Escherichia coli O157:H7 Infections
Srdjan Jelacic,1 Cheryl L. Wobbe,3 Daniel R. Boster,1
Marcia A. Ciol,4 Sandra L. Watkins,2 Phillip I. Tarr,1,5
and Ann E. Stapleton3
Divisions of 1Gastroenterology and 2Nephrology, Department
of Pediatrics, Children’s Hospital and Regional Medical
Center, 3Division of Allergy and Infectious Diseases, Department of
Medicine, and Departments of 4Rehabilitation Medicine
and 5Pediatrics and Microbiology, University of Washington
School of Medicine, Seattle
P1 and ABO antigens and bacterial stx genotypes might influence the risk of developing
hemolytic uremic syndrome (HUS) after Escherichia coli O157:H7 infections. We determined
ABO status and P1 antigen expression in 130 infected and 17 uninfected children, and we determined the stx genotype of the infecting isolate. P1 expression was weakly and directly related to
HUS risk (P ¼ .04), but this risk did not extend to the group with the greatest P1 expression. P1
expression remained constant as HUS evolved. The ABO frequency was similar in all groups.
These associations were not affected by the stx genotype. stx12/stx2+ E. coli O157:H7 strains
were more commonly associated with HUS than were stx1+/stx2+ strains (P ¼ .21), and 1 child
with HUS was infected with a rare stx1+/stx22 isolate. In the present study, ABO antigens and
stx genotypes were not major determinants of the outcome of E. coli O157:H7 infections, and
P1 expression did not protect against the development of HUS.
Escherichia coli O157:H7 cause diarrhea and bloody diarrhea. A subset of patients infected with E. coli O157:H7 develop
hemolytic uremic syndrome (HUS). HUS is a thrombogenic
microangiopathy believed to be precipitated by Shiga toxin
(Stx) absorbed from the gut that was produced by E. coli
O157:H7 [1]. HUS occurs 1 week after the onset of diarrhea
[2]. Presumably, Stx binds to vascular endothelial cells via globotriaosylceramide (Gb3), the glycosphingolipid receptor for
the Stx B subunit [3]. Subsequent toxin-mediated injury to the
endothelium causes microvascular injury that leads to hemolysis, thrombocytopenia, and extraintestinal injury, especially in
the kidneys.
Various erythrocyte antigens have been proposed to either
play a role in the development of Stx-mediated HUS or modulate
the severity of resulting HUS. For example, the P1 erythrocyte
Received 15 June 2001; revised 20 September 2001; electronically published
3 January 2002.
Presented in part: 4th International Symposium and Workshop on “Shiga
Toxin (Verocytotoxin)–Producing Escherichia coli Infections,” Kyoto,
Japan, 1 November 2000 (abstract 428), and the 12th Congress of the
International Pediatric Nephrology Association, Seattle, Washington, 4
September 2001 (abstract P375).
Informed consent was obtained from parents and informed assent was
obtained from subjects, if appropriate, prior to enrollment, according to
guidelines established by the Institutional Review Board of Children’s
Hospital and Regional Medical Center, Seattle.
Financial support: National Institutes of Health (grant 1RO1DK52081).
Reprints or correspondence: Dr. Ann E. Stapleton, Division of Allergy and
Infectious Diseases, Department of Medicine, Box 356523, University of
Washington School of Medicine, Seattle, WA 98195 ([email protected]).
The Journal of Infectious Diseases
2002;185:214–9
q 2002 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2002/18502-0010$02.00
antigen carries a terminal Gala1–4Gal residue capable of binding and neutralizing Stx1 and 2 in vitro [4], and it has been proposed that the binding of Stx by erythrocytes through the P1 antigen might prevent circulating toxin from binding to and injuring
endothelial cells. In one study, children with HUS whose erythrocytes expressed the P1 antigen had less-severe courses than
those whose erythrocytes were P1 negative, and children with
postdiarrheal HUS had a significantly lower expression of the
P1 blood group antigen on their erythrocytes, compared with
control children with renal disease but without a history of HUS
[5]. The theory that erythrocytes could express toxin-binding
glycolipids is also supported by the observation that the erythrocytes of children with HUS have lower Gb3 levels than do
erythrocytes of children with diarrhea secondary to Shiga toxin–
producing E. coli (STEC) who do not develop HUS [6]. However,
several subsequent epidemiological studies, mostly of children
infected during outbreaks, have failed to associate the expression of the P1 blood group antigen with diminished risk for
the development of HUS after infection with E. coli O157:H7
[7–10].
Stx also binds to a synthetic trisaccharide, Gala1–3Galb1–
4N-acetylglucosamine [11]. This trisaccharide epitope is expressed on cell surfaces of lower mammals but not on human
cells and is the target of the naturally occurring “anti-Gal” human
antibodies that have made xenotransplantation problematic [12,
13]. This trisaccharide is also identical to the B blood group antigen, minus a fucose linked to the subterminal galactose. Indeed,
Shimazu et al. [14] found that children with B blood type who
were infected during a large E. coli O157:H7 outbreak in Sakai,
Japan, in 1996 experienced a lower HUS rate than did infected
children with different blood types. The B blood group antigen
JID 2002;185 (15 January)
ABO and P1 Antigens and stx Genotype in Childhood HUS
was proposed to protect against the development of HUS in
children, possibly by binding circulating toxin.
The comparative virulences of Stx1 and 2 have also been proposed to affect disease outcome in E. coli O157:H7 infections.
The amino acid sequence of Stx1 is almost identical to that of
Stx produced by Shigella dysenteriae type 1, but it differs from
the sequence of Stx2 at 44% of its amino acid residues. Although
Stx1 and 2 contain some conserved sequences, they have different antigenic [15] and biological properties. In particular, Stx2
is considered to be the more potent of the 2 toxins in in vitro
[16] and in vivo studies [17]. Moreover, a significantly higher
proportion of children infected with isolates that contained Stx2
alone developed HUS than did the children who were infected
with isolates that contained Stx1 alone or Stx1 and 2 [18], and
STEC containing the gene encoding Stx2 were overrepresented
as the cause of severe illness in a study of 112 human isolates
[19]. Of interest, Donohue-Rolfe et al. [20] determined that
stx1+/stx2+ E. coli O157:H7 strains were less virulent in a gnotobiotic pig model of infection than a derivative from which the
stx1 gene was deleted.
We conducted a prospective analysis of risk factors for the
development of HUS after E. coli O157:H7 infections since
1997. Herein, we examine the relationship between the expression of ABO and P1 erythrocyte antigens in the host and the
stx genotype of the infecting strain on the development of HUS
in children infected with E. coli O157:H7.
Subjects and Methods
Subjects. The E. coli O157:H7– detection network has been
described elsewhere [21]. In brief, 46 participating laboratories in 4
states (Idaho, Oregon, Washington, and Wyoming) detect E. coli
O157:H7 by screening stools on sorbitol-MacConkey agar. Sorbitol
nonfermenting E. coli are then tested to determine whether they
express the O157 lipopolysaccharide antigen, by use of commercial
reagents. These laboratories then notify us of all patients ,10 years
old who have such presumptive E. coli O157:H7 infections. After
enrollment by the patients, blood is obtained for clinical and research purposes. In some cases, blood remaining from prior or subsequent clinical determinations was used for erythrocyte antigen
determinations. A standardized questionnaire is administered to
the caregivers of all subjects, to determine the timing of the onset of illness. Day 1 of illness is considered to be the first day of
diarrhea.
Case definitions. Infected children were considered to have
HUS if they had a hematocrit ,30%, with microangiopathic changes
on smears of peripheral blood; a platelet count ,150,000 cells/mm3;
and a serum creatinine concentration above the upper limit of
normal for age. Infected children were considered to have uncomplicated illness if they did not develop HUS. Children with
postdiarrheal HUS precipitated by E. coli O157:H7 who presented to the Children’s Hospital and Regional Medical Center
(CHRMC), Seattle, were also enrolled and studied if they had not
been enrolled during the pre-HUS stage.
215
Episodes of HUS were classified as oligoanuric if there was ,0.5
mL urine/kg body weight/h produced for >48 h during the HUS
phase and as nonoligoanuric if urine output exceeded this level.
Control erythrocytes were obtained from 17 children ,10 years
old who did not have inflammatory, hematologic, infectious, or nephrologic disorders and who underwent elective operations at the
CHRMC.
Specimen handling and erythrocyte antigen determinations.
Blood obtained by phlebotomy was evacuated into vacuum tubes.
After centrifugation, the clotted pellet was sent on ice to the laboratory of one of the authors (A.E.S.) for erythrocyte antigen determinations. The P1 and ABO antigens were determined by use of
macroscopic hemagglutination assays (Immucor), according to
standard blood-banking protocols [22]. Indeterminate P1 antigen
groupings were assessed microscopically. The person who performed the assays was not aware of the clinical status of the children
from whom the specimens were obtained. All specimens for the present study were obtained from phlebotomies performed before any
blood was transfused.
Toxin typing of infecting strains. stx genotypes were primarily
determined by polymerase chain reaction (PCR). Frozen isolates
were grown overnight in Luria Broth at 37 C; 45 mL of turbid culture
was added to 5 mL of 0.1% Triton X-100 (Sigma) and boiled for 20
min; and 5 mL of the boiled lysate was added to 45 mL of PCR mix.
The PCR mix contained dNTPs (200 mM of each nucleotide), 3 mM
MgCl2, 50 mM KCl, 10 mM Tris-HCl (pH 9.0 at 25 C), 0.1% Triton
X-100, 1.25 U DNA polymerase (Klenow fragment), and stx1 (50 CGC TTT GCT GAT TTT TCA CA-30 and 50 -GTA ACA TCG CTC
TTG CCA CA-30 ) or stx2 (50 -GTT CCG GAA TGC AAA TCA GT30 and 50 -CGG CGT CAT CGT ATA CAC AG-30 ) primers (50 pmol
of each primer). The PCR was performed in an iCycler (Bio-Rad
Laboratories) by use of the following conditions: 25 cycles of denaturation (1 min at 94 C), annealing (1 min at 55 C), and elongation
(2 min at 74 C) and 1 cycle of final elongation (7 min at 72 C). The
resulting respective 207- or 205-bp amplicons were visualized
after electrophoresis in 1% agarose– 0.5£ Tris-borate– EDTA
(TBE) gel and staining in ethidium bromide. E. coli O23:H15 strain
TB311 [23], E. coli O157:H7 strain 87-01 [18], E. coli O103:H6
strain TB154 [23], and E. coli O157:H7 strain 86-24 [24] were
the negative, stx1+/stx2+, stx1+/stx22, and stx12/stx2+ controls,
respectively.
For 2 isolates with unusual genotypes, Southern hybridizations
were performed to confirm the PCR genotyping. DNA was digested
with Bam HI, electrophoresed in 1% agarose–0.5£ TBE gel, and
transferred to a nylon membrane (Osmonics). The immobilized
DNA was probed under stringent conditions [25] with the inserts of
pJN37-19 and pNN111-19 to identify stx1 and stx2, respectively
[26]. E. coli O157:H7 strain 87-01 [18], E. coli O103:H6 strain
TB154 [23], and E. coli O157:H7 strain 86-24 [24] were the
stx1+/stx2+, stx1+/stx22, and stx12/stx2+ controls, respectively.
DNA from E. coli HB101 [25] was the negative control.
Statistics. The association between 2 categorical variables was
initially assessed by use of x2 tests. Some of the tables had ex-
pected values ,5, which required that we apply Fisher’s exact
test. For consistency, we reported only the Fisher’s exact test
results for all categorical data. Logistic regression was used to
assess the association between categorical variables (P1 and
216
Jelacic et al.
ABO blood group antigens and stx genotype) and the relative
risk that a child would develop HUS.
Results
Subjects enrolled. Between April 1997 and March 2001,
erythrocytes from 148 subjects were analyzed. These included
106 and 25 children with uncomplicated infection and HUS,
respectively, and 17 control subjects (table 1). The 25 children
with HUS included 7 children who were admitted to the CHRMC
at the time of HUS diagnosis but who were not enrolled prior to
the development of HUS. The isolate from 1 of the 106 subjects
with an uncomplicated infection had no evidence of stx genes
on either PCR or on Southern hybridization. This patient therefore was not included in these analyses, because of concern that
this infection might not have been related to the toxigenic properties of E. coli O157:H7 [27, 28].
P1 antigen and development of HUS. Children with uncomplicated infection, children with HUS, and healthy control subjects had similar distributions of the P1 blood group antigen
(P ¼ :33, Fisher’s exact test). One subject had a P1 level of 1+,
and this person was included in the 2+ group for purposes of
analysis. When the larger control group reported by Taylor et al.
[5] was substituted for comparison purposes for our healthy
control subjects, the distribution of the P1 blood group antigen
was not statistically different (P ¼ :23, Fisher’s exact test; figure
1). There was also no difference when the distribution of P1 antigen expression was compared with that of 2 larger control groups
recruited in Seattle for population-based studies [29]. These
patient groups (ages 18–40 years) had ethnic distributions similar to the group of infected children in this study.
ABO type and development of HUS. Children with uncomplicated infection, children with HUS, healthy control subjects,
and the Seattle population-based control subjects [29] had similar distributions of ABO blood group antigens (P ¼ :63, Fisher’s
exact test). Because we were interested in the association of the
presence of the B blood group antigen and the development of
HUS, children with blood types B and AB were compared with
children who expressed A and O blood group antigens. There
was no evidence of association between ABO blood group and
HUS development (P ¼ 1, Fisher’s exact test; figure 1).
P1 antigen expression as microangiopathy evolves. Blood
is subject to considerable shear stress as HUS evolves from gastrointestinal infection with E. coli O157:H7 [30], and it is possible that this shear stress might affect the expression of an
erythrocyte surface antigen, such as P1. Therefore, we determined the degree of expression of the P1 antigen in 4 subjects
whose erythrocytes were obtained on consecutive days between
the day of enrollment and the day that they developed HUS, to
determine whether evolving intravascular shear stress affects
the expression of this antigen. P1 antigen did not change despite
the evolution of profound hemolysis (figure 2).
JID 2002;185 (15 January)
stx genotype and development of HUS. stx12/stx2+ E. coli
O157:H7 were more frequently isolated than stx1+/stx2+ E.
coli O157:H7 from subjects with HUS than from those with
uncomplicated infection, but the difference was not statistically
significant (P ¼ :21, Fisher’s exact test; table 1). We also recovered an E. coli O157:H7 isolate that contained stx1 but not stx2
(genotype confirmed by Southern blot) from 1 patient with HUS.
This patient was excluded from further analysis because of the
possible bias introduced in statistical assessment by a single
observation.
Logistic regressions. Restricting the analysis to the infected
children, we studied the association among stx, ABO and P1
blood group antigens, and the risk of developing HUS. Models
with a single covariate were not statistically significant for stx
and ABO blood type (A þ O vs. B þ AB groups) (P ¼ :22 and
P ¼ :76, respectively). The model for P1 was marginally significant (P ¼ :052). A multivariate model was constructed that adjusted for all 3 covariates, in which P1 was entered after the
other 2 covariates were already in the model. In this model, the
addition of P1 was statistically significant (P ¼ :043). Table 2
shows the estimated odds ratio (OR) for each covariate and the
corresponding 95% confidence intervals (CIs). The estimated
OR is the risk that a person in a certain group develops HUS
compared with a person in the reference group. For example, a
person in the P1 ¼ 3þ group is 6.3 times more likely to develop
HUS than a person in the P1 ¼ 0 group, all other covariates
being the same for both persons. CIs that include the value 1
demonstrate that the estimate is not statistically different from
1. Therefore, only P1 ¼ 3þ would seem to have an effect on
the development of HUS. However, this category contains the
Table 1. Characteristics of study subjects and the stx genotype of
infecting isolates of Escherichia coli O157:H7.
Characteristic
Sex, male:female
Age, mean years ^ SD
Race or ethnic group
White
Hispanic
Black
Asian or Pacific Islander
Native American
stx Genotype of infecting
isolatesa
stx1+/stx22
stx12/stx2+
stx1+/stx2+
Children with
Healthy
control
subjects
(n = 17)
Uncomplicated
infection
(n = 106)
Hemolytic uremic
syndrome
(n = 25)
14:3
4.3 ^ 3.1
58:48
4.3 ^ 2.6
15:10
3.4 ^ 1.8
15 (88)
0
1 (6)
0
1 (6)
88 (83)
8 (7)
3 (3)
5 (5)
2 (2)
21 (84)
2 (8)
1 (4)
1 (4)
0
NA
NA
NA
0
30 (29)
75 (71)
1 (4)
10 (40)
14 (56)
NOTE. Data are no. (%) of patients, except where noted. NA, data not
available.
a
In 1 uncomplicated infection, the patient’s E. coli O157:H7 isolate had no
evidence of stx genes.
JID 2002;185 (15 January)
ABO and P1 Antigens and stx Genotype in Childhood HUS
217
Figure 1. Distribution of P1 (A) and ABO (B) blood group antigens among study groups of patients with uncomplicated Escherichia coli
O157:H7 infection (U), patients with hemolytic uremic syndrome (H), healthy control subjects (C), and “published” control subjects (P) [5, 22].
highest number of subjects from the HUS group (figure 1). We
also performed a logistic regression, adjusting for stx and ABO
blood group and adding P1 antigen in binary form (expressed
vs. not expressed). P1 was statistically significant in the model
(P ¼ :012), with an estimated OR of 4.445 (95% CI, 1.203–
16.423), which suggests that the presence of P1 on erythrocytes
is associated with the development of HUS by an infected host.
Discussion
Our demonstration that expression of the P1 blood group antigen is directly, although weakly, associated with the risk of
developing HUS after E. coli O157:H7 infection contrasts with
a previously published report [5]. Taylor et al. [5] and other
studies [7–10] generated a hypothesis that expression of specific
blood group antigens, such as ABO-B and P1, are protective
against HUS. However, several more-recent analyses [7–10]
have suggested that there is no association between P1 expression and infection outcome. The present study provides
stronger data against a protective effect of P1, because it represents an analysis of children infected by a diversity of E. coli
O157:H7 strains and not of children infected by a single strain,
as in outbreaks [7, 9]. Also, in comparing previous studies of
sporadic cases of HUS [5, 8, 10] in which the infecting isolates
were not genotyped, we were able to examine the interaction
between the P1 antigen as expressed by host erythrocytes, the
ABO blood group, and the stx genotype of the infecting strain.
It is possible that we overlooked some effect of genetic factors
on the expression of blood group antigen and HUS risk in our
mostly white population. However, it is noteworthy that race
218
Jelacic et al.
Figure 2.
Daily P1 blood group antigen expression and hemoglobin levels in 4 children with hemolytic uremic syndrome. ND, not determined.
was not a risk factor for the development of HUS in 3 previous
studies in the Pacific Northwest [21, 31, 32]. Furthermore, the
distribution of blood groups in the published control population
was similar to that found in an ethnically comparable population-based study from Seattle [29]. In addition, our sample
size of 130 infected subjects is larger than that of any previous
study that has examined P1 status and HUS risk. Finally, concerns that we studied these patients during acute illness, at a
point when erythrocytes were subject to considerable shear
stress, were addressed, because the P1 antigen was stable as
hemolysis accelerated.
The lack of a protective role of the B blood group antigen in
this diverse North American population is in contrast to the find-
Table 2. Multivariate analysis for the association among
stx genotype, ABO and P1 blood group antigens, and the
risk of developing hemolytic uremic syndrome in children
infected with Escherichia coli O157:H7.
Covariate
stx12/stx2+ genotypea
B + AB blood groupb
P1 blood groupc
P1 (1+ and 2+)
P1 (3+)
P1 (4+)
a
Reference group, stx1+/stx2+.
Reference group, A þ O.
c
Reference group, 0.
b
JID 2002;185 (15 January)
Odds ratio
(95% confidence interval)
1.986 (0.755–5.226)
1.195 (0.286–4.997)
3.282 (0.574–18.773)
6.285 (1.565–25.243)
3.087 (0.682–13.980)
ings of Shimazu et al. [14], who studied an exclusively Japanese
population involved in an outbreak. That study proposed that the
B blood group antigen might bind circulating toxin, thus preventing circulating toxin from binding to Gb3 on endothelial cells.
However, our data show that there is no statistical association
between the B antigen and protection against HUS, even when
the stx genotypes of infecting isolates are included in a multivariate model.
The relationship between the stx genotype of the infecting
strain and disease outcome is complex. Although we demonstrated a trend toward the stx12/stx2+ genotype being associated
more frequently with the development of HUS, the association
did not attain statistical significance. Also, it is worth noting that
1 of the children with HUS was infected with a rare stx1+/stx22
isolate, which is found in only 3% of clinical isolates [18]. To
our knowledge, an stx1+/stx22 E. coli O157:H7 isolate has been
recovered only once from a patient with HUS [33].
In summary, in this prospective study of childhood E. coli
O157:H7 infections caused by different infecting strains, we determined that the degree of expression of the P1 antigen on host
erythrocytes, the ABO blood type, and the presence of stx1 play
nonprotective roles with respect to the development of HUS. In
fact, our data suggest the possibility that the greater the expression of P1, the higher the frequency with which HUS occurs.
This effect was largely confined to the group positive at a 3+
level of P1 expression. It is, however, difficult to draw conclusions about this association, because this association was not
observed in the 4+ group. Therefore, in this North American
population of children, the erythrocyte antigens studied and the
JID 2002;185 (15 January)
ABO and P1 Antigens and stx Genotype in Childhood HUS
toxin genotypes observed in the infecting isolates are neither
major nor absolute risk factors for the development of HUS
after E. coli O157:H7 infection.
Acknowledgments
We thank Katheryn Green and Rachelle Whitman for expert
secretarial assistance in the preparation of this manuscript and the
participating patients and families, microbiologists and other laboratory technicians, nurses, and physicians for providing specimens
and data used in this study.
References
1. Nakao H, Takeda T. Escherichia coli Shiga toxin. J Nat Toxins 2000; 9:
299– 313.
2. Tarr PI, Neill MA, Clausen CR, Watkins SL, Christie DL, Hickman RO.
Escherichia coli O157:H7 and the hemolytic uremic syndrome: importance of early cultures in establishing the etiology. J Infect Dis 1990;162:
553– 6.
3. Jacewicz M, Clausen H, Nudelman E, Donohue-Rolfe A, Keusch G.
Pathogenesis of shigella diarrhea. XI. Isolation of a shigella toxin–
binding glycolipid from rabbit jejunum and HeLa cells and its identification as globotriaosylceramide. J Exp Med 1986; 163:1391–404.
4. Bitzan M, Richardson S, Huang C, Boyd B, Petric M, Karmali MA. Evidence that verotoxins (Shiga-like toxins) from Escherichia coli bind to
P blood group antigens of human erythrocytes in vitro. Infect Immun
1994;62:3337–47.
5. Taylor CM, Milford DV, Rose PE, Roy TC, Rowe B. The expression of
blood group P1 in post-enteropathic haemolytic uraemic syndrome.
Pediatr Nephrol 1990; 4:59– 61.
6. Newburg DS, Chaturvedi P, Lopez EL, Devoto S, Fayad A, Cleary TG. Susceptibility to hemolytic-uremic syndrome relates to erythrocyte glycosphingolipid patterns. J Infect Dis 1993; 168:476–9.
7. Ashida A, Matsui K, Chizaki T, et al. Erythrocyte P1 group antigen expression in VTEC-associated hemolytic uremic syndrome. Clin Nephrol
1999;51:73–6.
8. Robson WL, Leung AK, Bowen T, Brant R, Ching E. The P1 blood group
and the severity of diarrhea-associated hemolytic uremic syndrome. Clin
Nephrol 1994; 42:288–90.
9. Orr PH, Dong V, Schroeder ML, Ogborn MR. P1 blood group antigen
expression and epidemic hemolytic uremic syndrome. Pediatr Nephrol
1995; 9:612–3.
10. Green D, Murphy W, Uttley W. Haemolytic uraemic syndrome: prognostic
factors. Clin Lab Haematol 2000; 22:11–4.
11. Armstrong GD, Fodor E, Vanmaele R. Investigation of Shiga-like toxin
binding to chemically synthesized oligosaccharide sequences. J Infect
Dis 1991; 164:1160–7.
12. Galili UWL, LaTemple DC, Radic MZ. The natural anti-Gal antibody. Subcell Biochem 1999; 32:79–106.
13. Parker W, Stitzenberg KB, Yu PB, Pratt VS, et al. Biophysical characteristics of anti-Gal(a)1-3Gal IgM binding to cell surfaces: implications for
xenotransplantation. Transplantation 2001; 71:440–6.
14. Shimazu T, Shimaoka M, Taenaka N, Hasegawa T. Does blood type B protect against haemolytic uraemic syndrome? An analysis of the 1996
Sakai outbreak of Escherichia coli O157:H7 (VTEC O157) infection. J
Infect Dis 2000; 41:45– 9.
219
15. Donohue-Rolfe A, Acheson DW, Kane AV, Keusch GT. Purification of
Shiga toxin and Shiga-like toxins I and II by receptor analog affinity chromatography with immobilized P1 glycoprotein and production of crossreactive monoclonal antibodies. Infect Immun 1989; 57:3888–93.
16. Louise CB, Obrig TG. Specific interaction of Escherichia coli O157:H7–
derived Shiga-like toxin II with human renal endothelial cells. J Infect
Dis 1995; 172:1397–401.
17. Tesh VL, Burris JA, Owens JW, et al. Comparison of the relative toxicities
of Shiga-like toxins type I and type II for mice. Infect Immun 1993; 61:
3392– 402.
18. Ostroff SM, Tarr PI, Neill MA, Lewis JH, Hargrett-Bean N, Kobayashi JM.
Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coli O157:H7 infections. J Infect Dis 1989; 160:994– 8.
19. Boerlin P, McEwen SA, Boerlin-Petzold F, Wilson JB, Johnson RP, Gyles CL.
Associations between virulence factors of Shiga toxin–producing Escherichia coli and disease in humans. J Clin Microbiol 1999;37:497–503.
20. Donohue-Rolfe A, Kondova I, Oswald S, Hutto D, Tzipori S. Escherichia
coli O157:H7 strains that express Shiga toxin (Stx) 2 alone are more neurotropic for gnotobiotic piglets than are isotypes producing only Stx1 or
both Stx1 and Stx2. J Infect Dis 2000; 181:1825–9.
21. Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI. The risk of hemolytic
uremic syndrome after antibiotic treatment of Escherichia coli O157:H7
infections. N Engl J Med 2000; 342:1930– 36.
22. Walker RH. Technical manual. 10th ed. Arlington, VA: American Association of Blood Banks, 1990.
23. Bokete TN, Whittam TS, Wilson RA, et al. Genetic and phenotypic analysis
of Escherichia coli with enteropathogenic characteristics isolated from
Seattle children. J Infect Dis 1997; 175:1382–9.
24. Tarr PI, Bilge SS, Vary J, et al. Iha: a novel Escherichia coli O157:H7 adherence-conferring molecule encoded on a recently acquired chromosomal
island of conserved structure. Infect Immun 2000; 68:1400–7.
25. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1982.
26. Newland JW, Neill RJ. DNA probes for Shiga-like toxins I and II and for
toxin-converting bacteriophages. J Clin Microbiol 1988; 26:1292–7.
27. Schmidt H, Scheef J, Huppertz H, Frosch M, Karch H. Escherichia coli
O157:H7 and O157:H(2) strains that do not produce Shiga toxin: phenotypic and genetic characterization of isolates associated with diarrhea and
hemolytic-uremic syndrome. J Clin Microbiol 1999; 37:3491–6.
28. Allerberger F, Dierich M, Gruber-Moesenbacher U, et al. Nontoxigenic
sorbitol-fermenting Escherichia coli O157:H2 associated with a family
outbreak of diarrhoea. Wien Klin Wochenschr 2000; 112:846–50.
29. Scholes D, Hooten TM, Roberts PL, Stapleton AE, Gupta K, Stamm WE.
Risk factors for recurrent urinary tract infection in young women. J Infect
Dis 2000; 182:1177–82.
30. Tsai HM, Chandler WL, Sarode R, et al. Escherichia coli O157:H7– associated hemolytic uremic syndrome plasma has normal von Willebrand factor–cleaving proteaseinase activity but increased von Willebrand factor
proteolysis. Pediatr Res 2001; 49:653–9.
31. Tarr PI, Hickman RO. Hemolytic uremic syndrome epidemiology: a population-based study in King County, Washington, 1971 to 1980. Pediatrics
1987; 80:41–5.
32. Rogers MFRG, Alexander SR, DiLiberti JH, Foster L, Schonberger LB,
Hurwitz ES. A population-based study of hemolytic-uremic syndrome in
Oregon, 1979–1982. Am J Epidemiol 1986; 123:137–42.
33. Kim YB, Okuda J, Matsumoto C, et al. Isolation of an Escherichia coli
O157:H7 strain producing Shiga toxin 1 but not Shiga toxin 2 from a
patient with hemolytic uremic syndrome in Korea. FEMS Microbiol Lett
1998; 166:43–8.